On 17/08/2017 18:50, Radim Krčmář wrote:
> 2017-08-17 13:14+0200, David Hildenbrand:
>>> atomic_set(&kvm->online_vcpus, 0);
>>> mutex_unlock(&kvm->lock);
>>> diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
>>> index c8df733eed41..eb9fb5b493ac 100644
>>> --- a/include/linux/
2017-08-17 13:14+0200, David Hildenbrand:
> > atomic_set(&kvm->online_vcpus, 0);
> > mutex_unlock(&kvm->lock);
> > diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
> > index c8df733eed41..eb9fb5b493ac 100644
> > --- a/include/linux/kvm_host.h
> > +++ b/include/linux/kvm_host
> atomic_set(&kvm->online_vcpus, 0);
> mutex_unlock(&kvm->lock);
> diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
> index c8df733eed41..eb9fb5b493ac 100644
> --- a/include/linux/kvm_host.h
> +++ b/include/linux/kvm_host.h
> @@ -386,12 +386,17 @@ struct kvm_memslots {
On Wed, 16 Aug 2017 21:40:37 +0200
Radim Krčmář wrote:
> This is a prototype with many TODO comments to give a better idea of
> what would be needed.
Just a very superficial reading...
>
> The main missing piece a rework of every kvm_for_each_vcpu() into a less
> inefficient loop, but RCU read
This is a prototype with many TODO comments to give a better idea of
what would be needed.
The main missing piece a rework of every kvm_for_each_vcpu() into a less
inefficient loop, but RCU readers cannot block, so the rewrite cannot be
scripted. Is there a more suitable protection scheme?
I di
5 matches
Mail list logo